Breeding Biology of Atlantic Population Canada Geese in Nunavik, Northern Québec RICHARD C

ARCTIC VOL. 66, NO. 3 (SEPTEMBER 2013) P. 301 – 311

Breeding Biology of Atlantic Population Canada Geese in Nunavik, Northern Québec RICHARD C. COTTER,1,2 R. JOHN HUGHES,3 PETER MAY, 4 PAULUSI NOVALINGA,5 JIMMY JOHANNES,5 LARRY J. HINDMAN6 and PAUL I. PADDING7

(Received 12 May 2011; accepted in revised form 30 January 2013)

ABSTRACT. The Atlantic population of Canada Geese (Branta canadensis interior) experienced a sharp decline in numbers in the late 1980s. Management agencies in Canada and the United States responded by implementing several measures, notably closing sport hunting seasons for a number of years in most Atlantic Flyway states and provinces and funding a research project to study the nesting ecology and the factors affecting productivity of this goose population. In this paper we present the nesting phenology and breeding biology of Atlantic population Canada Geese on their tundra nesting grounds in Nunavik, Québec, specifically on a primary study area (32.8 km2) along Hudson Bay (1997 – 2003) and on several secondary sites (most smaller than 1 km2) distributed along the coastal lowlands of Hudson Bay and Ungava Bay (1996 – 2005). In the late 1990s the population rebounded, with strong increases in the population of breeding pairs and the density of nests between 1996 and 2001, followed by stabilization of both variables from 2001 to 2005. As a result, there was a near doubling in productivity index (the number of goslings produced per km2) on the primary study area, from 17.9 in 1997 to 32.0 in 2003. Geese start laying eggs soon after snow disappears from their nesting grounds. On the primary study area, for all years pooled, mean clutch initiation date, clutch size, hatching date, and Mayfield nesting success were 27 May, 4.54 eggs, 26 June, and 67.3%, respectively. Among secondary sites along both Hudson Bay and Ungava Bay, we found a correlation between the annual average daily temperature for 4 – 24 May (the period leading up to egg-laying) and the annual mean clutch initiation date: higher temperatures were correlated with earlier initiation. Similarly, an earlier annual mean clutch initiation date was correlated with large mean clutch size. Key words: Canada Goose, Branta canadensis interior, Atlantic population, breeding biology, nesting success, productivity, Nunavik, Ungava Peninsula, Québec

RÉSUMÉ. La population de l’Atlantique de la bernache du Canada (Branta canadensis interior) a subi un déclin prononcé à la fin des années 1980. Les organismes chargés de la gestion de la faune au Canada et aux États-Unis ont réagi en mettant en place plusieurs mesures. Ils ont notamment fermé la saison de chasse sportive dans la plupart des États et provinces de la voie migratoire de la population de l’Atlantique pendant un certain nombre d’années et financé un projet de recherche sur l’écologie de la nidification et sur les facteurs limitant la productivité de cette population. Dans le présent article, nous décrivons la phénologie de la nidification et la biologie de la reproduction de la population de l’Atlantique de la bernache du Canada dans son aire de nidification de la toundra du Nunavik (Québec), plus précisément dans une aire d’étude principale (32,8 km2) établie en bordure de la baie d’Hudson (1997 – 2003) et dans plusieurs sites secondaires de superficie plus modeste (dans la plupart des cas, inférieure à 1 km2) disséminés dans les basses-terres côtières de la baie d’Hudson et de la baie d’Ungava (1996 – 2005). À la fin des années 1990, la population a rebondi, le nombre de couples nicheurs et la densité des nids augmentant fortement entre 1996 et 2001 pour ensuite se stabiliser de 2001 à 2005. Par suite de ce redressement, le nombre d’oisons produits par kilomètre carré a presque doublé (indice de productivité) dans l’aire d’étude principale, passant de 17,9 en 1997 à 32,0 en 2003. La ponte débute peu après la disparition de la neige dans les sites de nidification. Dans l’aire d’étude principale, nous avons enregistré les valeurs moyennes suivantes (années regroupées) : début de la période de ponte : 27 mai; taille des couvées : 4,54 œufs; premières éclosions : 26 juin; succès de la nidification en mai : 67,3 %. Aux sites secondaires, en bordure de la baie d’Hudson et de la baie d’Ungava, nous avons relevé une corrélation entre la température journalière moyenne annuelle pour la période du 4 au 24 mai (période précédant le début de la période de ponte) et la date annuelle moyenne du début de la ponte: les températures plus élevées donnaient une corrélation avec un début de période de ponte plus hâtif. De même, une date annuelle moyenne plus hâtive du début de la ponte a été corrélée avec une grande taille moyenne des couvées.

1 Environment Canada, Canadian Wildlife Service, 801-1550 d’Estimauville Avenue, Québec, Québec G1J 0C3, Canada 2 Corresponding author: [email protected] 3 Environment Canada, Canadian Wildlife Service, 335 River Road, Ottawa, Ontario K1A 0H3, Canada 4 Nunavik Research Centre, Makivik Corporation, PO Box 179, Kuujjuaq, Québec J0M 1C0, Canada 5 Nunavik Hunting, Fishing, and Trapping Association, PO Box 879, Kuujjuaq, Québec J0M 1C0, Canada 6 Maryland Department of Natural Resources, 828B Airpax Road, Suite 500, Cambridge, Maryland 21613, USA 7 Division of Migratory Bird Management, United States Fish and Wildlife Service, 11510 American Holly Drive, Laurel, Maryland 20708, USA © Her Majesty the Queen in Right of Canada. Administered by the Arctic Institute of North America 302 • R.C. COTTER et al.

Mots clés : bernache du Canada, Branta canadensis interior, population de l’Atlantique, biologie de la reproduction, succès de la nidification, productivité, Nunavik, péninsule d’Ungava, Québec

Révisé pour la revue Arctic par Nicole Giguère.

Ivujivik Salluit

Nunavik Kangiqsujuaq Ungava Akulivik Peninsula Quaqtaq Québec Hudson Puvirnituq Bay Kangirsuk Ungava Bay Primary Aupaluk Study Area Inukjuak Tasiujaq Kangiqsualujjuaq

Kuujjuaq Newfoun and Labrador

Umiujaq dland

100 km

FIG. 1. Location of the Canada Goose primary study area and the secondary sites (small circles) within the Hudson Bay and Ungava Bay lowlands on the Ungava Peninsula in northern Québec.

INTRODUCTION of the breeding grounds on the Ungava Peninsula in Nuna- vik, in 1988, counted 118 000 (± 15 144 [SE]) nesting pairs Up to the 1980s, the Atlantic population (AP) of Canada (Malecki and Trost, 1990), but an aerial survey in 1995 Geese (Branta canadensis interior) was the most abundant counted only 33 995 (± 3045) pairs (Harvey and Rodrigue, Canada Goose population in North America, with a mid- 2012). This decline prompted wildlife management agen- winter estimate of nearly one million birds in 1981 (Hind- cies in Canada and the United States to close sport hunt- man and Ferrigno, 1990) and an annual sport harvest in ing seasons throughout most of the Atlantic Flyway in the Atlantic Flyway estimated at more than 400 000 birds 1995 and to produce the Action Plan for the Atlantic Popu- (Malecki and Trost, 1990). This population nests entirely lation of Canada Geese (Atlantic Flyway Council, 1996), in Québec, with more than 80% breeding in Nunavik, which detailed objectives and strategies for the recovery of Québec’s Arctic region north of 55˚ N, and the remainder the population, including monitoring and research needs. nesting in the taiga and northern boreal forest as far south A key objective was to document recruitment parameters as 48˚ N. Its major wintering areas are the Delmarva Penin- such as nesting phenology, clutch size, and nesting suc- sula of Maryland and Delaware and portions of New York, cess. The Action Plan therefore recommended a multi-year Pennsylvania, New Jersey, and Virginia (Dickson, 2000; study of the breeding ecology (primary study area) and Atlantic Flyway Council, 2008). In the late 1980s and early the implementation of an annual monitoring program to 1990s, this population experienced a dramatic decline in measure breeding effort and success at key locations (sec- size as a result of high harvest rates and a number of breed- ondary sites) on the Ungava Peninsula. The monitoring pro- ing seasons with poor productivity. The first aerial survey gram was established in 1996, and the research study was BREEDING BIOLOGY OF CANADA GEESE • 303 conducted from 1997 to 2003. This paper presents the key surveyed in only two years, and few nests were found there. reproductive parameters from the breeding ecology study On average, the Ungava Bay sites ranged in size between and the first 10 years of the nesting monitoring program, 0.25 km2 and 2.46 km2. 1996 to 2005.

METHODS STUDY AREA Primary Study Area This study was carried out on the Ungava Peninsula in the Nunavik territory of northern Québec (Fig. 1) from Inuit from Puvirnituq, the nearest village (60 km dis- 1996 to 2005. This tundra region lies within the Arctic Eco- tant) to the primary study area, begin observing Canada climatic Province (Canada Committee on Ecological Land Geese in the region each year, usually in early to mid May. Classification, 1989), specifically the Southern Arctic Eco- Each year the field crew timed their arrival to coincide zone, an area characterized by herb-lichen tundra, a very with the start of egg laying, but in some years they had to cold, dry climate, and continuous permafrost (Wiken et al., wait to begin nest searches because of delayed snowmelt 1996). The highest densities of breeding pairs of AP Can- (Table 1). As soon as it was apparent that some geese had ada Geese are found on this peninsula in two regions: the initiated nesting, the field crew (five or six people, depend- coastal lowlands of eastern Hudson Bay and the coastal ing on the year) began to search for nests: members walked lowlands of southwestern Ungava Bay (Malecki and Trost, approximately 50 m apart and searched the study area sys- 1990; Harvey and Rodrigue, 2012), primarily between the tematically, with emphasis on a close search of the edges Inuit communities of Inukjuak and Akulivik on Hudson of wetlands and ponds. Three to five days were required to Bay and between Kuujjuaq and Kangirsuk on Ungava Bay. search the entire study area, depending on the weather and We therefore chose the location of our study areas—a large the amount of snow cover. When a nest was found, it was primary study area (1997 – 2003 only) and smaller second- assigned a unique number, a 1 m stake (painted fluorescent ary study sites along Hudson Bay and Ungava Bay—from orange) was placed at a distance of 25 m (to help relocate these two regions (Fig. 1). it on subsequent visits), and each egg in the nest was num- The primary study area, 32.8 km2 in size, was situ- bered. The following information was also recorded: geo- ated along the Polemond River at 59˚ 31.5′ N, 77˚ 36.1′ W, graphic coordinates (using a hand-held Global Positioning approximately 10 km inland from the Hudson Bay coast System [GPS], Garmin models 48, 76; NAD27 datum); date and 60 km south of Puvirnituq. It is characterized by and time; status of nest when found (laying, incubating, lichen-heath tundra (approximately 65% of total area), lakes hatching, abandoned, destroyed, or unknown); the number (22%), wet sedge meadows (11%), and ponds and streams of eggs; the status of each egg (warm or cold, intact or bro- (2%) (Cadieux et al., 2005). Lichens, dwarf birch (Betula ken, in or out of the nest); the mass (nearest g), length (near- glandulosa), mountain cranberry (Vaccinium vitis-idaea), est 0.1 mm), and width (nearest 0.1 mm) of each egg; the and crowberry (Empetrum nigrum) are the dominant plant shortest distance from the nest to water (m); and the three cover in the lichen-heath tundra, whereas wet-sedge mead- most common plant species within a 1 m radius of the nest, ows are composed mostly of mosses, water sedge (Carex as well as the approximate cover of each species. Further- aquatilis), and cottongrass (Eriophorum angustifolium). more, all nests were classified into two broad habitat cat- Edges of most ponds are dominated by sedges (Carex spp.) egories (dry and wet) based on proximity to a waterbody. and cottongrass, whereas along lakes willow (Salix lanata) Dry nests were more than 10 m away from a waterbody is the dominant plant species (Cadieux et al., 2005). larger than 0.0021 ha (e.g., 3 m × 7 m), and wet nests were The secondary study sites were located in two stretches within 10 m of such a waterbody. of coastal lowlands, one along Hudson Bay and the other After the initial visit, we revisited all nests approxi- along Ungava Bay, each approximately 180 km long. Along mately every three days until either no new eggs were Hudson Bay there were seven secondary sites located found in the nest or the nest was depredated or abandoned. between the Mariet River (59˚ 9′ N) in the south and the For new eggs, the same information was recorded as dur- Korak River (60˚ 46′ N) in the north. These sites ranged ing the first visit; if no new eggs had been laid, we assumed in size from 0.35 km2 to 0.64 km2 and all were surveyed laying was completed, the clutch size (CS) for the nest was in every year except 1996, when only five sites were sur- recorded, and the nest was not revisited until mid-incuba- veyed. Along Ungava Bay there were 10 secondary sites; tion, at which time it was visited every 2 – 3 days, with the however, no more than six were surveyed in any one year, penultimate visit timed to occur with the start of hatch- and only six sites were surveyed in five or more years. With ing. The last visit occurred 1 – 2 days after pipping began one exception, these sites were located between the Whale in order to web-tag each gosling and to record clutch size River (Rivière à la Baleine), located approximately 40 km at hatch (CSH), the number of goslings leaving the nest east of Kuujjuaq in the south and Aupaluk in the north. (GLN), and hatching success (HS = GLN/CS). The exception was a site located farther east along Ungava Clutch initiation date was defined as the date the first egg Bay, about 30 km northwest of Kangiqsualujjuaq; it was was laid and was calculated for each nest by subtracting a 304 • R.C. COTTER et al.

TABLE 1. General weather conditions and timing of the field crew arrival and searches for Canada Goose nests at the primary study area and at the secondary study sites along Hudson Bay and Ungava Bay in northern Québec, 1996 – 2005.

Nest search date3 1 Average temperature (˚C) May 4 – 24 Arrival at primary study area First egg laid2 Primary area Hudson Bay Ungava Bay Year Puvirnituq Kuujjuaq % snow cover Date (May) (May) (May) (June) (June)

1996 -5.4 -3.2 14 – 16 7 – 11 1997 0.2 2.0 10 22 23 24 6 – 8 6 – 9 1998 2.4 4.3 5 21 17 22 9 – 11 3 – 6 1999 0.9 2.7 25 20 17 21 5 – 8 7 – 10 2000 -1.4 -0.1 95 20 24 28 10 – 14 5 – 8 2001 3.5 5.6 5 20 18 21 9 – 13 4 – 9 2002 -4.6 -2.4 95 18 54 94 16 – 20 10 – 14 2003 0.0 5.5 75 17 22 24 7 – 12 5 – 11 2004 -3.7 -1.1 12 – 15 11 – 17 2005 0.2 3.9 10 – 13 7 – 8

1 Average of the mean daily temperatures (average of daily maximum and daily minimum) recorded at the Puvirnituq Airport (on Hudson Bay) and Kuujjuaq Airport (on Ungava Bay) (available from http://climate.weatheroffice.gc.ca/). 2 Earliest clutch initiated on the primary study area. 3 First day of nest searches on the primary study area and the Hudson Bay and Ungava Bay secondary study sites. 4 The month is June (for 2002). number equal to total clutch size minus one from the date have died before leaving the nest were excluded. A factor on which incubation began (i.e., the date the last egg was that could bias gosling survival is total brood loss (when no laid). Eggs are laid at intervals of approximately 35 hours members of a brood survive), which is impossible to dis- (see Mowbray et al., 2002); therefore, for large clutches tinguish from marked broods simply not being encoun- (four or more eggs) an egg was not subtracted from the total tered during banding. An indication of annual total brood clutch size. Hatch date was defined as the date on which loss differences is obtained by calculating the proportion all eggs in the nest had begun pipping (star-pip or hole-pip of marked broods (% broods recaptured) for which one or stage), and length of incubation was calculated by subtract- more goslings were recaptured. ing the start of incubation from the hatch date. An annual index of population productivity, the produc- The fate of every nest found was recorded as either suc- tivity index (PI) or the number of goslings produced per cessful (at least one egg hatched), abandoned (clutch still km2, was calculated as follows: intact but eggs cold), destroyed (i.e., predation causing complete absence of eggs or presence of broken egg shells PI = nest density • Mayfield nest success • without separated egg membranes), observer destroyed (one goslings leaving per nest • gosling survival nest in 1997 and three nests in 1998), infertile, or unknown. Both apparent nesting success (the percentage of nests initi- Each year after hatching had occurred, a crew of two ated that hatched at least one gosling) and Mayfield nesting or three observers visited a sector within the study area success (Mayfield, 1961; Johnson, 1979) were calculated. each day over a four-hour period (0700 – 1100, 1100 – 1500, Nests found after they had failed were included in calcu- 1500 – 1900) and recorded the location and size of all indi- lating apparent nesting success but excluded for the May- vidual family broods observed. We considered broods of field method, which requires a nest to be under observation 10 or more goslings as having goslings from more than (i.e., exposure) for at least one day to be included in the cal- one brood and therefore excluded them from the analyses. culations. Furthermore, for the Mayfield method, if a nest These observations continued until banding commenced, was observed to have failed between two visits, the date of at which time (for most years) goslings were four to six failure was assumed to be halfway between the two visits weeks of age. The distances moved by broods, specifically (Klett and Johnson, 1982). web-tagged goslings that were captured during the band- At hatch, we fitted each gosling with a uniquely num- ing drives in late July to mid August, were determined by bered web-tag (size 1, National Band and Tag Co.). Marked measuring the distance from the goslings’ nests to their goslings and their parents were recaptured four to six weeks banding (catch) locations. later during the annual banding drives. For every brood There is evidence from several Arctic sites that small from which at least one marked gosling was recaptured, we mammal populations fluctuate widely between years and calculated gosling survival from hatch to banding as the that these fluctuations can have a profound effect on pre- proportion of goslings that had left the nest (where status dation rates on goose eggs (Bêty et al., 2001; Wilson and of leaving the nest was either confirmed, probably left the Bromley, 2001). To monitor small mammal populations nest, or unknown) and were subsequently recaptured dur- (primarily Ungava lemming Dicrostonyx hudsonius), ing banding. Goslings that were marked but were known to we followed the protocol established for the Northwest BREEDING BIOLOGY OF CANADA GEESE • 305

Territories and Nunavut (see Carrière, 1999). In 1997, four (0.003 g/cm3) for AP Canada Geese nesting in the taiga of 250 m transects were set up in two different habitat types— north-central Québec. This method was used for all sites two transects each in lowland and upland habitats. Museum in all years, except at Hudson Bay sites in 2000, where the Special snap traps were placed along each transect at 10 m stage of incubation was determined by floating eggs and intervals. Each year 100 traps were set and checked once nest age was calculated following Walter and Rusch (1997). a day, between 0800 and 1000, for 10 days. Trapped indi- The density index method computes the start of incubation viduals were identified to species using Lupien (2002) and for individual eggs within each nest by subtracting the day Desrosiers et al. (2002). The same trapping transects were of incubation (i.e., DAYS, except for Hudson Bay sites in used each year, and they were set up in the second half of 2000) from the date the egg was measured (presumably the July. date the nest was found). Taking the average of the eggs in each clutch, we obtained the start of incubation date for Secondary Study Sites each nest. We obtained the clutch initiation date (the start of egg laying) for each nest by subtracting a number of days Secondary nest sites along Hudson Bay and Ungava Bay equal to total eggs in the nest minus one from the date on were first visited during early incubation, the timing deter- which incubation began. As for nests on the primary study mined from nesting phenology information from the pri- area, we did not subtract an egg from the total for clutches mary study area or from observations of local Inuit. At this of four or more eggs. We estimated hatch date by adding 26 visit to a site, two pairs of observers systematically ground- days to the first day of incubation. For Hudson Bay sites in searched the entire area, recorded the location of each nest 2000, following Walter and Rusch’s (1997) method for large found and marked it with a stake, numbered each nest and Canada Geese, we first calculated the clutch age as equal any eggs it contained, and recorded the same information as to 4.333 × stage – 2.167, where stage equals the stage of for nests in the primary study area. development of the yolk. Hatch date = date found + (length Each site was visited a second time after hatching. of incubation clutch age), where length of incubation is 26 To economize on helicopter travel, this visit was car- days. With respect to stage of yolk development, Walter and ried out during the banding operations in late July to mid Rusch (1997) used the six float stages described by Wester- August. Each nest was revisited and its fate was recorded skov (1950) for Ring-necked Pheasants (Phasianus colchi- as either successful (i.e., egg caps or membranes were cus). In our study, we grouped these six into three stages. present), or abandoned (all eggs were still in the nest), or Clutch initiation date was then calculated by subtracting predator-destroyed (absence of egg shells or presence of (length of incubation + (clutch size –1))—or (length of incu- egg fragments with membrane attached). The area of each bation + clutch size) for clutches of four or more eggs— secondary site was calculated using the mapping software from the hatch date. ArcView with one of two methods. For sites along Hudson Bay, ArcView was used to map the boundary of each site, Statistical Analyses which remained constant among years, and then calculate the surface area. The boundaries of Ungava Bay sites, on Annual means (± standard errors, SE) for each repro- the other hand, were not fixed, and therefore we plotted the ductive variable were calculated for the primary site and locations of all nests and ArcView calculated the surface all secondary sites using the following nest criteria: nest area using the nests on the perimeter (Minimum Convex density and nest fate (i.e., percentage of total nests initiated Polygon). that were either successful, abandoned, destroyed, or other) Since most nests were found during incubation, we calcu- used all nests that were found; clutch initiation date and lated clutch initiation date by first determining, for each egg clutch size for the primary study area used only nests found in the clutch, the number of days of incubation (DAYS) at during egg-laying; clutch initiation date for the secondary the time of the first visit. Basing our method on the principle sites used all active nests found; clutch size for the sec- that eggs gradually lose mass over the course of incubation, ondary sites used only nests found during incubation; and we calculated DAYS using an index of egg density (DI) and clutch size at hatch, number of goslings leaving the nest, a regression of density on days, where for each egg: hatching success, and hatching date were determined only for successful nests on the primary study area (whether DI = (mass / length • width2) • 1000, and found during egg-laying or incubation). The annual means DAYS = (DI of fresh eggs – DI of measured egg) / of each variable for Hudson Bay and Ungava Bay, with one daily rate of change in density exception, were calculated by pooling nests from across all = (0.5551 – DI) / 0.003 sites. The exception was nest density, for which the annual mean was the mean of the nest densities of the individual Since we had neither a measure of DI for fresh eggs nor sites along either Hudson Bay or Ungava Bay (and there- the daily rate of change in density for tundra-nesting AP fore sample size was equal to the number of sites surveyed). Canada Geese, we used Cooper’s (1978) DI value (0.5551 The standard errors for each nest fate (successful, aban- g/cm3) obtained for a Canada Goose population in Man- doned, destroyed, and other) were based on the binomial itoba and Hughes et al.’s (2000) rate of change in density distribution: 306 • R.C. COTTER et al.

SE = √(pq)/n 5.4,61) 49.7 77.0 27.5 60.7 63.6 22.7 56.9 12.5 20.3 60.2 5.3, 59) 5.0, 80) 3.7, 181) 3.1, 185) 4.6, 113) 4.6, 116) where, for example, p = the proportion of nests abandoned, 3.2, 225) ± ± ± ± ± ± ± ( (± 9.4, 28) 9.4, (± ± ( (± 8.5, 16) 8.5, (± ( ( ( ( ( q = 1 – p, and n = total number of nests. Ungava Bay

For Hudson Bay and Ungava Bay secondary sites, site ( (fixed variable) and year (random variable) were tested in a 77.7 49.7 75.1 89.8 83.8 85.5 78.0 72.4 82.9 68.3 2.7,127) 130) 3.7, 159) 4.0, 167) 3.6, 2.1, 297) 2.1, 2.6, 273) 2.8, 187) two-factor mixed-model analysis of variance (PROC GLM; 2.3, 228) ± ± ± ± ± ± 58) 5.9, (± ± ± ( 123) 3.7, (± ( ( ( ( ( ( ( SAS Institute, 2004) for clutch initiation date, clutch size, Hudson Bay

and apparent nest success. Correlations between sets of two Nesting success (%) variables were analyzed by calculating Pearson product-

75.7 55.1 23.0 86.8 82.5 84.2 64.3 1.8,348) 675) 1.7, 639) 1.4, 2.3, 423) 2.2, 383) moment correlation coefficients (r) (PROC CORR; SAS 2.8, 325) Primary ± ± ± ± ± ± ( ( ( ( ( Institute, 2004). The significance level was set at α < 0.05 2.2, (± 292) for all tests. (

4.31 3.47 3.42 3.69 3.89 4.07 4.36 4.23 4.34 4.22 0.15, 81) 33) 0.23, 142) 0.11, 187) 0.10, RESULTS 136) 0.10, ± ± 0.13, 91) ± 0.13, ± 9) 0.39, (± ± (± 0.20,55) (± ± ± ( ( (± 0.15, 67) 0.15, (± ( 0.30, 22) (± ( Ungava Bay ( (

Nest Density and Site Characteristics 3.11 4.15 3.21 3.53 3.95 4.55 4.52 4.38 4.66 Over the 10 years of this study, 5898 Canada Goose nests 4.25 120) 0.11, 0.07, 251) 0.08, 141) 0.07, 176) 0.07, 0.07, 280) 0.07, 0.08, 208) ± ± ± ± 55) 0.17, (± ± Clutch size ± 0.10,121) (± ( ( ( Hudson Bay 108) 0.10, (± ( 128) 0.10, (± ( were found: 3085 on the primary study area, 1749 on Hud- ( son Bay secondary sites, and 1064 on Ungava Bay secondary sites. On the primary study area (32.8 km2), nest density

2 2 3.63 4.49 4.65 5.29 4.52 4.98 4.05 0.12, 73) increased from 8.9 nests/km in 1997 to 20.6 nests/km in 0.10,102) 0.11, 111) 111) 0.11, 0.07, 162) 162) 0.07, 0.07, 322) 322) 0.07, 0.06, 248) Primary ± ± ± ± ± ± ( ( ( (± 0.09, 113) 113) 0.09, (± ( 2003 (Table 2), a 131% increase. Nest densities were con- ( siderably higher on the smaller secondary sites than on the

primary study area, two to five times higher in any given ( year (Table 2). Between 1997 and 2003, the secondary sites 82) 0.7, 1.0, 27) 1.0, 168) 0.9, 142) 0.5, along Hudson Bay experienced a level of increase in nest 137) 0.9, 11 June 11 01 June01 30 May30 20 May 22 May 24 May 20 May 24 May 24 May 07 June ± ± 0.6, 91) ± 0.6, ± 0.8,55) (± ± ± ± ( 19) 1.0, (± ( 12) 1.4, (± 72) 0.4, (± ( ( ( ( density (121%) similar to that of the primary study area, Ungava Bay from 37.7 nests/km2 to 83.4 nests/km2. From 1996 to 2005, nest density on these sites increased by 249%, which is 121) 0.6, 153) 0.5, 251) 0.6, 176) 0.4, 280) 0.5, very close to the 241% increase in the size of the breeding 208) 0.3, 19 May 19 01 June01 01 June01 23 May 2005. 28 May 28 May 22 May 05 June 05 June 09 June ± ± ± ± ± ± 55) 0.5, (± 0.4,114) (± ( ( ( – 109) 0.4, (± ( 125) 0.3, (± population recorded by Harvey and Rodrigue (2012) on the Hudson Bay

Ungava Peninsula over this 10-year period (from 51 466 ± Clutch initiation date 4789 [SE] pairs to 175 679 ± 16 737 pairs). Both nest density , 178)

0.2, 99) 0.1, 322) 322) 0.1, 0.2, 199) 0.2, 199) and the size of the breeding population increased strongly 0.2, 284) 31 May 31 21 May 21 11 June 11 29 May29 23 May 28 May 24 May Primary ± ± 0.2,110) ± 0.2,110) ± ± 0.2 ± ± ( ( ( ( ( (± 0.2, 127) (± up to 2001 but then stabilized in the following years. On the ( Ungava Bay secondary sites, nest densities increased each year from 1996 to 1999 but afterwards fluctuated consider- ably (Table 2). ( 3.5 (61) (59) (16) (28) (80) 67.1 19.1 19.4 21.2 14.2 40.7 43.3 38.4 24.5 (181) (113) (116) ( More Canada Goose nests were found in wet habitat than (185) ) 2 in dry habitat. Wet habitat represented 67% (2073 of 3085 Ungava Bay nests) on the primary study area, 85% (1026 of 1207 nests) on Hudson Bay secondary sites, and 60% (429 of 716 nests) (58) 37.7 71.1 51.2 38.7 49.2 83.4 25.5 56.4 88.9 40.2 (123) (187) (225) (273) on Ungava Bay secondary sites. Wet or dry habitat was not (297)

recorded for 157 of 873 nests (18%) of Ungava Bay sites. Hudson Bay

The mean distance of nests to water was 7.4 m (± 23.7 [SD], Nest density (n/km n = 3078 nests) on the primary study area, 4.1 m (± 7.7, n

9.9 8.9 11.7 19.5 10.6 12.9 20.6 (228) (639) (130) (423) (159) (383) (675) (292) (167) (325) = 1186 nests) on Hudson Bay secondary sites, and 8.2 m (127) (348) (± 15.3, n = 649 nests) on Ungava Bay secondary sites. At Primary all three areas, the most common plant providing nest cover was dwarf birch; this shrub was ranked first (i.e., providing the most cover) at 54% of all nests at the primary study area (all years and habitat types pooled), 38% at Hudson Bay secondary sites, and 28% at Ungava Bay secondary sites. The next most common plants bordering goose nests were TABLE 2. Mean (± SE, n) nest initiation density, clutch SE, 2.n) MeanTABLE (± date, size, clutch and apparent nesting success Canada of Geese the for primary study area (Primary) and the for secondary study sites along Hudson and in Bay Bay Ungava northern Québec, 1996 Year 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 BREEDING BIOLOGY OF CANADA GEESE • 307 crowberry at the primary study area, grasses (Graminae range spp.) at Hudson Bay, and willow and sedge at Ungava Bay. median

10th 90th Nesting Phenology, Egg Size, Clutch Size, and Hatching percentile percentile Success All Years On the primary study area, 80% (10th to 90th percentile) 2003 of nests over all years were initiated (i.e., first egg laid) dur- 2002 ing a 15-day period from 21 May to 4 June (Fig. 2), and the 2001 overall mean date of initiation was 27 May (± 0.2, n = 1319 2000 nests). Among years we observed a three-week difference between the earliest year (80% initiation: 18 – 24 May 1998) 1999 and the latest year (9 – 13 June 2002) (Fig. 2). In any one 1998 year, 80% of all nests were initiated over a five- to nine-day 1997 period. This period was longer for the secondary sites: 9 to 05/0905/14 05/1905/24 05/2906/03 06/0806/13 06/1806/23 24 days for Hudson Bay and 9 to 26 days for Ungava Bay. Date (mm/dd) At both secondary areas, there was a non-significant site effect (Hudson Bay: F = 2.04, df = 6, p = 0.08; Ungava Bay: FIG. 2. Phenology of clutch initiation by Canada Geese at the primary study F = 1.15, df = 9, p = 0.36), but there were significant effects area in northern Québec, 1997–2003. of year (Hudson Bay: F = 35.94, df = 9, p < 0.01; Ungava Bay: F = 19.20, df = 9, p < 0.01) and interaction effects eggs for secondary sites on Hudson Bay and Ungava Bay (Hudson Bay: F = 4.86, df = 52, p < 0.01; Ungava Bay: F = (Table 2). Clutches with 3 – 6 eggs represented 93% (1052 1.61, df = 20, p = 0.04). Overall mean clutch initiation date of 1131 nests) of all clutches in the primary study area, was 27 May (± 0.4, n = 805 nests, 95% CI = 26 – 27 May) at 89% (1409 of 1588 nests) of secondary sites at Hudson Bay, Ungava Bay and 30 May (± 0.2, n = 1592 nests, 95% CI = and 83% (686 of 823 nests) of those at Ungava Bay. For the 29 – 30 May) at Hudson Bay. primary study area only, the mean clutch size (CS) across Annual mean clutch initiation date was negatively cor- years was 4.54 eggs, clutch size at hatch (CSH) was 4.08 related with the mean of the average daily temperatures eggs, and number of goslings leaving the nest (GLN) was (i.e., average of the daily minimum and maximum tempera- 3.66 goslings (Table 3). The mean GLN was 19.4% smaller tures) at both Hudson Bay and Ungava Bay (temperatures than the mean CS, and this difference provides an indica- recorded at Puvirnituq and Kuujjuaq airports, respectively) tion of egg loss during incubation and gosling loss during for the three-week period leading up to egg laying (4 – 24 hatching. This loss ranged from 16.0% (2001) to 26.4% May): primary study area (r = 0.94, n = 7 years), Hudson (2000). Among secondary sites, across years and sites the Bay secondary sites (r = 0.91, n = 10 years), and Ungava Bay mean clutch size was 4.09 eggs (± 0.03, n = 1588 nests, 95% secondary sites (r = 0.90, n = 10 years). With this mean tem- CI = 4.02 – 4.15) at Hudson Bay and 4.03 eggs (± 0.05, n = perature (4 – 24 May), the following model can be used to 823 nests, 95% CI = 3.94 – 4.13) at Ungava Bay. There was calculate a predicted mean clutch initiation date at the onset a significant site effect at Hudson Bay (F = 3.99, df = 6, p = of nesting. For Hudson Bay, clutch initiation date = ‒2.1043 0.002), but not at Ungava Bay (F = 0.49, df = 9, p = 0.87). • temperature + 27.9376, and for Ungava Bay, clutch initia- At Hudson Bay, however, Tukey’s Studentized Range test tion date = ‒2.1194 • temperature + 31.2453. showed that the mean clutch sizes for six of the seven sites The mean length of incubation on the primary study area were not significantly different. At both areas, there was a was 26 days in every year except 2003, when it was 25 days. significant year effect (Hudson Bay: F = 27.64, df = 9,p < Across years the mean date of hatching was 26 June, with 0.01; Ungava Bay: F = 3.08, df = 9, p = 0.001), and there was annual means ranging from 20 June to 10 July (Table 3). a significant interaction effect at Hudson Bay (F = 1.45, df Overall (years and primary and secondary sites pooled), = 52, p = 0.02) but not at Ungava Bay (F = 1.46, df = 20, eggs had a mean length of 83.4 mm (± 0.03, n = 18 992 p = 0.09). At all three areas, annual mean clutch size was eggs), width of 56.3 mm (± 0.01, n = 18 992 eggs), and mass positively correlated with the annual mean temperature for of 140.0 g (± 0.09, n = 18 992 eggs). For the Hudson Bay 4 – 24 May (primary study area: r = 0.77, n = 7 years; Hud- region (i.e., pooling the primary and secondary sites), the son Bay: r = 0.90, n = 10 years; Ungava Bay: r = 0.77, n = mean length, width, and mass were 83.2 mm (± 0.03, n = 10 years) and negatively correlated with the annual mean 15 711 eggs), 56.1 mm (± 0.01, n = 15 711 eggs), and 139.3 g clutch initiation date (primary study area: r = 0.83, n = 7 (± 0.10, n = 15 711 eggs), respectively, and for Ungava Bay years; Hudson Bay: r = 0.85, n = 10 years; Ungava Bay: r = the same means were 84.2 mm (± 0.06, n = 3281 eggs), 0.90, n = 10 years). Hatching success (GLN/CS) of clutches 56.9 mm (± 0.03, n = 3281 eggs), and 143.8 g (± 0.25, n = at the primary study area, across years, was 80% (± 0.7 3281 eggs), respectively. [SE], n = 923 nests), with annual means ranging from 72% The annual means for clutch size varied between 3.6 and to 88%. 5.3 eggs for the primary study area and between 3.1 and 4.7 308 • R.C. COTTER et al.

TABLE 3. Means (± SE) of reproductive parameters of Canada Geese nesting at the primary study area in northern Québec, 1997 – 2003. Sample sizes are shown in parentheses.

Parameter 1997 1998 1999 2000 2001 2002 2003 Years pooled

Clutch size (CS) 4.98 ± 0.09 5.29 ± 0.10 4.65 ± 0.07 4.05 ± 0.11 4.49 ± 0.07 3.63 ± 0.12 4.52 ± 0.06 4.54 ± 0.03 (n = 113) (n = 102) (n = 162) (n = 111) (n = 322) (n = 73) (n = 248) (n = 1131)

Clutch size at hatch (CSH) 4.47 ± 0.08 4.72 ± 0.07 3.75 ± 0.08 3.45 ± 0.12 4.30 ± 0.05 3.03 ± 0.08 3.93 ± 0.06 4.08 ± 0.03 (n = 227) (n = 301) (n = 253) (n = 87) (n = 534) (n = 177) (n = 498) (n = 2077)

Goslings leaving nest (GLN) 4.16 ± 0.09 4.26 ± 0.07 3.45 ± 0.08 2.98 ± 0.12 3.77 ± 0.05 2.71 ± 0.08 3.51 ± 0.06 3.66 ± 0.03 (n = 231) (n = 298) (n = 248) (n = 84) (n = 532) (n = 174) (n = 501) (n = 2068)

Hatching date 29 June ± 0.2 20 June ± 0.1 22 June ± 0.2 30 June ± 0.3 24 June ± 0.1 10 July ± 0.2 27 June ± 0.1 26 June ± 0.1 (n = 220) (n = 293) (n = 245) (n = 75) (n = 505) (n = 170) (n = 475) (n = 1983)

Brood size1 3.91 ± 0.18 3.92 ± 0.08 3.39 ± 0.10 3.07 ± 0.50 4.19 ± 0.13 2.71 ± 0.09 3.33 ± 0.06 3.53 ± 0.04 (n = 198) (n = 416) (n = 193) (n = 14) (n = 161) (n = 224) (n = 679) (n = 1885)

Brood movements (m) 1554 ± 69 4667 ± 241 7809 ± 546 6384 ± 1684 9359 ± 537 7756 ± 1034 11914 ± 772 7237 ± 239 (nest to banding site) (n = 215) (n = 297) (n = 209) (n = 16) (n = 282) (n = 80) (n = 265) (n = 1364)

Mayfield nesting success (%)2 81.9 89.4 61.3 20.4 81.7 50.8 77.6 67.3 (n = 35; (n = 27; (n = 124; (n = 239; (n = 92; (n = 105; (n = 102; (n = 724; 76.6 – 87.6) 85.6 – 93.3) 56.2 – 67.0) 16.6 – 25.1) 78.3 – 85.2) 44.5 – 57.9) 73.8 – 81.6) 65.4 – 69.4)

Successful nests (%) 82.5 ± 2.2 86.8 ± 1.8 64.3 ± 2.3 23.0 ± 2.1 84.2 ± 1.4 55.1 ± 2.8 75.7 ± 1.7 69.1 ± 0.8 (Apparent nesting success) (n = 292) (n = 348) (n = 423) (n = 383) (n = 639) (n = 325) (n = 675) (n = 3085)

Abandoned nests (%) 0.7 ± 0.5 0.3 ± 0.3 0.7 ± 0.4 3.4 ± 0.9 0.3 ± 0.2 1.8 ± 0.7 1.3 ± 0.4 1.2 ± 0.2 (n = 292) (n = 348) (n = 423) (n = 383) (n = 639) (n = 325) (n = 675) (n = 3085)

Destroyed nests (%) 15.4 ± 2.1 9.8 ± 1.6 34.5 ± 2.3 73.4 ± 2.3 15.5 ± 1.4 43.1 ± 2.7 23.0 ± 1.6 29.2 ± 0.8 (n = 292) (n = 348) (n = 423) (n = 383) (n = 639) (n = 325) (n = 675) (n = 3085)

Other (%)3 1.4 ± 0.7 3.2 ± 0.9 0.5 ± 0.3 0.3 ± 0.3 0.0 ± 0.0 0.0 ± 0.0 0.0 ± 0.0 0.6 ± 0.1 (n = 292) (n = 348) (n = 423) (n = 383) (n = 639) (n = 325) (n = 675) (n = 3085)

1 Mean brood size of all broods observed on study area, from hatching up to banding (approximately 4 – 6 weeks of age). 2 Sample size and 95% confidence intervals in parentheses. 3 From 1997 to 2003, other causes were observer destroyed (n = 4 nests), infertility (n = 10 nests), and unknown (n = 4 nests).

Nesting Success and Productivity n = 7 years; secondary sites: r = 0.46, n = 10 years). The mean Mayfield nesting success for the primary study The fate was determined for 99.9% (3081 of 3085 nests) area for all years pooled was 67%, and annual means of all nests found in the primary study area, 99.5% (1740 ranged from a low of 20% in 2000 to a high of 89% in of 1749 nests) at Hudson Bay secondary sites, and 95.5% 1998 (Table 3). The proportion of nests abandoned never (1016 of 1064 nests) at Ungava Bay secondary sites. Pool- exceeded 4% in any one year at any of the areas (primary ing years, apparent nesting success (the proportion of nests and secondary), except at Ungava Bay in 2000, when 24% initiated that hatched at least one gosling) was 69.1% for the (19 of 80 nests) were abandoned. The likely cause was a primary study area, 77.2% (± 1.0, n = 1749 nests, 95% CI late-season snowfall in mid-June. = 75.2 – 79.2) for Hudson Bay, and 47.8% (± 1.5, n = 1064 Gosling survival (at approximately four weeks of age) nests, 95% CI = 44.8 – 50.8) for Ungava Bay (Table 3). at the primary study area on the Polemond River was rela- There was no site effect at either Hudson Bay (F = 1.35, df tively stable from year to year, ranging from 57% to 63% = 6, p = 0.25) or Ungava Bay (F = 1.13, df = 9, p = 0.39), and in five of the seven years (Table 4). In the other two years, at Ungava Bay there was no year effect either (F = 2.07, df it was much lower: 42% in 2000 and 48% in 2001. In 2000, = 9, p = 0.08). However, there was a year effect at Hudson the primary study area not only had a late spring snowmelt, Bay (F = 5.81, df = 9, p < 0.01) and a significant interaction but also received the most July precipitation (total rain- effect at both Hudson Bay (F = 2.42, df = 52, p < 0.01) and fall recorded at the camp) of any year, 80.9 mm. In con- Ungava Bay (F = 3.90, df = 20, p < 0.01). Apparent nest- trast, 2001 had the lowest July rainfall, only 18.7 mm. The ing success was positively correlated with the mean tem- 2000 nesting season, in addition to having the lowest gos- perature during 4 – 24 May for Ungava Bay (r = 0.88, n = 10 ling survival of any year, also had the lowest nesting suc- years), but not for either the primary study area or second- cess and the lowest mean number of goslings leaving the ary sites along Hudson Bay (primary study area: r = 0.62, nest (Table 3). The number of goslings produced per km2, BREEDING BIOLOGY OF CANADA GEESE • 309

TABLE 4. Gosling survival and productivity index of Canada Geese nesting at the primary study area in northern Québec, 1997 – 2003.

Parameter 1997 1998 1999 2000 2001 2002 2003

Broods marked1 211 278 237 75 472 166 455 Broods recaptured2 89 124 93 10 159 45 142 % broods recaptured 42.2 44.6 39.2 13.3 33.7 27.1 31.2 Goslings marked 842 1166 804 225 1799 465 1558 Goslings recaptured 215 295 210 16 284 81 277 % gosling survival3 58.9 57.1 63.4 41.7 47.5 58.4 57.0 (± 3.2) (± 2.5) (± 2.9) (± 8.6) (± 2.1) (± 4.3) (± 2.4) Productivity index (goslings/km2) 17.9 23.1 17.3 3.04 28.5 8.0 32.0

1 Number of broods in which one or more goslings were marked with a tag. 2 Number of broods from which one or more marked goslings were recaptured at banding. 3 Standard error in parentheses and sample size is the number of marked broods recaptured. 4 The percentage of marked broods recaptured at banding was much lower than average, suggesting that the rate of total brood loss may have been high. or productivity index (PI), ranged from a low of only 3.0 and then it tripled in the next two years (2002: 32 per 1013 goslings/km2 in 2000 to a high of 32.0 in 2003 (Table 4) trap-nights; 2003: 30 per 1030 trap-nights). Of the small and was not significantly correlated with the annual mean mammals caught, 92% were Ungava lemmings, and the temperature of 4 – 24 May (r = 0.69, n = 7 years), clutch ini- remainder were either Gapper’s red-backed voles (Clethri- tiation date (r = ‒0.59, n = 7 years), or clutch size (r = 0.56, onomys gapperi) or meadow voles (Microtus pennsylvani- n = 7 years). cus). There was no correlation between the number of small The mean size of broods observed on the primary study mammals captured per 100 trap-nights and the annual area (years pooled) declined over the summer periods: mean Mayfield nesting success (r = 0.06, n = 6 years), 3.58 goslings at hatch to one week old, 3.54 goslings at two annual mean brood size (r = -0.58, n = 6 years), or annual weeks, 3.44 goslings at four weeks, and 3.00 goslings at six productivity index (r = 0.07, n = 6 years). weeks. The distance moved by goslings from the time they were web-tagged at hatch to the time they were captured at four to six weeks of age during banding drives varied from DISCUSSION as little as 56 m to 61 km, with an overall average across years of 7237 m (Table 3). The median distance, pooling After experiencing a sharp decline in the size of the years, was 4.0 km, and for single years it ranged from 1.4 breeding population in the late 1980s and early 1990s, from to 7.3 km. an estimated 118 000 pairs in 1988 (Malecki and Trost, 1990) to 33 995 pairs in 1995 (Harvey and Rodrigue, 2102), Nest Predation and Small Mammal Abundance Atlantic Population Canada Geese rebounded strongly over the next six years. This rebound was due to two main fac- The percentage of nests destroyed each year by predators: 1) the measures implemented by the Atlantic Flyway tors ranged from 10% to 73% at the primary study area Council in 1995, notably the closure of sport hunting sea- (Table 3), 10% to 48% at the Hudson Bay secondary sites, sons in most Atlantic Flyway states and provinces between and 16% to 88% at the Ungava Bay secondary sites. Across 1995 and 1999 (Atlantic Flyway Council, 2008), and 2) years and sites, the mean rate was 29.2% at the primary three consecutive breeding seasons of good productivity study area (Table 3), 22.1% (± 1.0, n = 1749 nests) at Hud- that occurred while sport hunting seasons were closed (in son Bay secondary sites, and 45.8% (± 1.5, n = 1064 nests) 1997 – 99). From the historic low in 1995 to 2001, the size at Ungava Bay secondary sites. At the primary study area, of the breeding population more than quadrupled to 135 196 there was no correlation between the annual rate of nest pairs (Harvey and Rodrigue, 2012). This increase was predation (%) and either the annual mean clutch initiation matched by a strong increase in nest density, as illustrated date (r = 0.51, n = 7 years) or annual nest density (r = ‒0.22, by an increase of 179% on the Hudson Bay secondary sites n = 7 years). Nevertheless, the two years with latest clutch between 1996 and 2001. During that period, only 2000 was initiation, 2000 and 2002, experienced the highest preda- a poor production year, primarily as a result of low nest suc- tion rates—pooling these two years, the average was 59.4% cess and lower than average clutch size. (± 1.8, n = 708 nests) compared with only 20.2% (± 0.8, n = The number of breeding pairs began to stabilize after 2377 nests) for the other years (pooled). 2001, as the sport harvest had resumed in Atlantic Flyway The number of small mammals captured annually, states and provinces. Interestingly, even though the number per 100 trap-nights, was less than 1.0 from 1998 to 2001 of breeding pairs did not change significantly from 2001 to (1998: 8 per 938 trap-nights; 1999: 5 per 951 trap-nights; 2005, annual production varied substantially; the produc- 2000: 5 per 1043 trap-nights; 2001: 5 per 1046 trap-nights), tivity index in our primary study area was 28.5 goslings/ 310 • R.C. COTTER et al. km2 in 2001, 8.0 in 2002, and 32.0 in 2003. This variation observe any correlation between small mammal abundance was due to large annual changes in nest density, suggest- and either nesting success or the productivity index. ing that once the population had recovered sufficiently, the main determinant of annual productivity was the propensity of breeding pairs to nest. The highest nest densities ACKNOWLEDGEMENTS occurred during the last year of the study (2003 in the primary study area, 2005 on secondary sites), so nesting hab- Financial support was provided by the James Bay Energy itat did not appear to be limited. Thus, propensity to nest Corporation (JBEC) and Arctic Goose Joint Venture (project may have been influenced more by the conditions of nest- #59) in 1996 and thereafter by the Arctic Goose Joint Venture, ing habitats when the birds arrived on the breeding grounds the U.S. Fish and Wildlife Service, and the Atlantic Flyway than by the number of breeding pairs present. Council and with additional assistance from Ducks Unlimited In most years, AP Canada Geese arrive on their breed- Inc., the Blue Mountain Conservation Fund (Pennsylvania), the ing grounds in northern Québec during the first two weeks Lancaster County Chapter-Waterfowl USA, the Lehigh Valley of May (P. Novalinga, J. Johannes, and P. May, pers. obs.), a Chapter Safari Club, and the Susquehanna River Wetlands period when open ground appears as the snow melts and nest Trust. The Wildlife Management Institute facilitated financial sites become available. In our study, we found that the mean administration. Austin Reed and Jack Hughes of the Canadian temperature during this period, specifically from 4 to 24 Wildlife Service and the Nunavik Hunting, Fishing and Trapping May, is an important influence on key breeding parameters. Association, in particular Paulusi Novalinga and Jimmy There were significant correlations between this temperature Johannes, were instrumental in getting the study off the ground. and clutch initiation date (negative) and clutch size (positive), Bill Doidge of Makivik Corporation also showed strong support and a weak positive correlation (p = 0.08) with productivity and provided considerable logistical help out of Kuujjuaq, as index (PI, number goslings per km2). In their study of factors did Québec’s Ministère des Ressources naturelles et de la Faune affecting clutch size in Canada Geese nesting on the west- in Kuujjuaq. Each year Aliva Tulugak in Puvirnituq assisted ern shore of Hudson Bay, MacInnes and Dunn (1988) deter- greatly with logistics, and his help and friendship were greatly mined that the interval between arrival and clutch initiation appreciated. was more important than the lateness of the season per se, as This study would not have been possible without the dedicated the geese during that interval had to use some of their accu- work in the field of everyone who participated: Alix Gordon, mulated energy reserves for maintenance—reserves that Anne Lagacé, Austin Reed, Barry Ford, Bryan Swift, Caroline otherwise could have been used for egg production. Rochon, Catherine Poussart, Chesley Mesher, David Sausville, As has been documented in other populations of Arc- Eric Reed, Francis St Pierre, Gary Costanzo, Geoff Kline, George tic-nesting geese (Mickelson, 1975; Leafloor et al., 2000; Haas, George Timko, Gérald Picard, Guillaume Tremblay, H. Lepage et al., 2000; Ely et al., 2008), in any given year AP Heusmann, Jack Hughes, Jean-Pierre Tremblay, Joël Poirier, Canada Geese were relatively synchronized in nest initia- John Dunn, Josée Lefebvre, Kathy Dickson, Larry Hindman, tion, with 80% of clutches on the primary study site initi- Marie-Christine Cadieux, Mark Dionne, Mélanie Bédard, Natalie ated during a 5- to 9-day period. In most years, the timing Hamel, Paul Castelli, Peter May, Pierre Brousseau, Richard of nest initiation of these tundra-nesting AP geese is similar Cotter, Salvatore Cozzolino, Sandy Suppa, Simeonie Qamaluk, to that of taiga-nesting AP geese, for which Hughes et al. Simon Bachand, Stacey Jarema, Stéphane Forest, Stéphane (2000) obtained annual means (1993 – 96) that ranged from Menu, Steven Heerkens, Tagralik Partridge, Ted Nichols, and 18 to 25 May on their study area near the Laforge Reservoir Vanessa Richard. Special thanks to Paul Castelli (NJ) for finding (650 km southeast of our primary study area). the U.S. volunteers to participate in the fieldwork. We would Some studies of Arctic nesting geese have shown that like to thank the Canadian Coast Guard and Nunavik Rotors, nesting success and productivity are directly related to in particular the helicopter pilots, who ferried us safely to our small mammal abundance. In years of low small mammal secondary study sites along both Hudson Bay and Ungava Bay. abundance, predators, especially generalist predators such as Arctic fox (Vulpes lagopus) and Herring Gull (Larus argentatus) (Pierotti and Good, 1994; Angerbjörn et al., REFERENCES 1999; Bêty et al., 2001), both of which are common along both Hudson and Ungava Bays (P. Novalinga, J. Johannes, Angerbjörn, A., Tannerfeldt, M., and Erlinge, S. 1999. Predator- and P. May, pers. obs.), will switch from small mammals to prey relationships: Arctic foxes and lemmings. Journal of goose eggs and nests (Bêty et al., 2001). Across the Cana- Animal Ecology 68(1):34 – 49. dian Arctic, small mammal populations fluctuate in a three- Atlantic Flyway Council. 1996. Action plan for the Atlantic to four-year cycle (Krebs et al., 2002). However, this cycle population of Canada Geese. Laurel, Maryland: Atlantic was not observed on our primary study area along eastern Flyway Council. Hudson Bay: for four consecutive years, the population was ———. 2008. A management plan for the Atlantic population very low, and then it increased to a level more than four of Canada Geese. Laurel, Maryland: Atlantic Flyway Council times as high for each of the next two years. We did not Migratory Game Bird Technical Section. http://www.dnr.state. md.us/wildlife/Hunt_Trap/waterfowl/geese/AP_CanadaGeese. asp. BREEDING BIOLOGY OF CANADA GEESE • 311

Bêty, J., Gauthier, G., Giroux, J.-F., and Korpimäki, E. 2001. Are Krebs, C.J., Kenney, A.J., Gilbert, S., Danell, K., Angerbjörn, A., goose nesting success and lemming cycles linked? Interplay Erlinge, S., Bromley, R.G., Shank, C., and Carrière, S. 2002. between nest density and predators. Oikos 93(3):388 – 400. Synchrony in lemming and vole populations in the Canadian Cadieux, M.-C., Gauthier, G., and Hughes, R.J. 2005. Feeding Arctic. Canadian Journal of Zoology 80(8):1323 – 1333. ecology of Canada Geese (Branta canadensis interior) Leafloor, J.O., Hill, M.R.J., Rusch, D.H., Abraham, K.F., and Ross, in sub-Arctic inland tundra during brood-rearing. Auk R.K. 2000. Nesting ecology and gosling survival of Canada 122(1):144 – 157. Geese on Akimiski Island, Nunavut, Canada. In: Dickson, Canada Committee on Ecological Land Classification. 1989. K.M., ed. Towards conservation of the diversity of Canada Ecoclimatic regions of Canada. Ecological Land Classification Geese (Branta canadensis). Occasional Paper No. 103. Ottawa: Series No. 23. Ottawa: Canadian Wildlife Service, Environment Canadian Wildlife Service, Environment Canada. 109 – 116. Canada. 118 p. Lepage, D., Gauthier, G., and Menu, S. 2000. Reproductive Carrière, S. 1999. Small mammal survey in the Northwest: Report consequences of egg-laying decisions in Snow Geese. Journal 1998. Manuscript Report No. 115. Yellowknife: Department of of Animal Ecology 69(3):414 – 427. Resources, Wildlife and Economic Development, Government Lupien, G. 2002. Recueil photographique des caractéristiques of the Northwest Territories. 22 p. morphologiques servant à l’identification des micromammifères Cooper, J.A. 1978. The history and breeding biology of the Canada du Québec. Jonquière: Société de la faune et des parcs du Geese of Marshy Point, Manitoba. Wildlife Monograph No. 61. Québec. 26 p. 87 p. MacInnes, C.D., and Dunn, E.H. 1988. Components of clutch size Desrosiers, N., Morin, R., and Jutras, J. 2002. Atlas des variation in Arctic-nesting Canada Geese. Condor 90(1):83 – 89. micromammifères du Québec. Société de la faune et des parcs Malecki, R.A., and Trost, R.E. 1990. A breeding ground survey du Québec, Direction du développement de la faune. Québec. of Atlantic Flyway Canada Geese, Branta canadensis, in 92 p. northern Québec. Canadian Field-Naturalist 104:575 – 578. Dickson, K.M. 2000. The diversity of Canada Geese. In: Dickson, Mayfield, H. 1961. Nesting success calculated from exposure. K.M., ed. Towards conservation of the diversity of Canada Wilson Bulletin 73:255 – 261. Geese (Branta canadensis). Occasional Paper No. 103. Ottawa: Mickelson, P.G. 1975. Breeding biology of Cackling Geese and Canadian Wildlife Service, Environment Canada. 11 – 24. associated species on the Yukon-Kuskokwim Delta, Alaska. Ely, C.R., Pearce, J.M., and Ruess, R.W. 2008. Nesting biology of Wildlife Monograph 45. 35 p. Lesser Canada Geese, Branta canadensis parvipes, along the Mowbray, T.B., Ely, C.R., Sedinger, J.S., and Trost, R.E. 2002. Tanana River, Alaska. Canadian Field-Naturalist 122(1):29 – 33. Canada Goose (Branta canadensis). In: Poole, A., and Gill, F., Harvey, W.F., and Rodrigue, J. 2012. A breeding pair survey of eds. The birds of North America, No. 682. Philadelphia: The Canada Geese in northern Québec – 2012. Unpubl. report. Birds of North America, Inc. Maryland, Department of Natural Resources and Canadian Pierotti, R.J., and Good, T.P. 1994. Herring Gull (Larus Wildlife Service, Québec region. http://www.dnr.state.md.us/ argentatus). In: Poole, A., and Gill, F., eds. The birds of North wildlife/Hunt_Trap/waterfowl/geese/AP_CanadaGeese.asp. America, No. 124. Philadelphia: The Academy of Natural Hindman, L.J., and Ferrigno, F. 1990. Atlantic flyway goose Sciences; Washington, D.C.: The American Ornithologists’ populations: Status and management. Transactions of the Union. North American Wildlife and Natural Resources Conference SAS Institute Inc. 2004. SAS/STAT 9.1 user’s guide. Cary, North 55:293 – 311. Carolina: SAS Institute, Inc. Hughes, R.J., Reed, A., Rancourt, L., and Bergeron, R. 2000. Walter, S.E., and Rusch, D.H. 1997. Accuracy of egg flotation Breeding ecology of Canada Geese near the Laforge-1 in determining age of Canada Goose nests. Wildlife Society hydroelectric reservoir in north-central Québec. In: Dickson, Bulletin 25(4):854 – 857. K.M., ed. Towards conservation of the diversity of Canada Westerskov, K. 1950. Methods for determining the age of game Geese (Branta canadensis). Occasional Paper No. 103. Ottawa: bird eggs. Journal of Wildlife Management 14(1):56 – 67. Canadian Wildlife Service, Environment Canada. 99 – 107. Wiken, E., Gauthier, D., Marshall, I., Lawton, K., and Hirvonen, Johnson, D.H. 1979. Estimating nest success: The Mayfield H. 1996. A perspective on Canada’s ecosystems: An overview method and an alternative. Auk 96(4):651 – 661. of the terrestrial and marine ecozones. Occasional Paper No. Klett, A.T., and Johnson, D.H. 1982. Variability in nest survival 14. Ottawa: Canadian Council on Ecological Areas. 95 p. rates and implications to nesting studies. Auk 99:77 – 87. Wilson, D.J., and Bromley, R.G. 2001. Functional and numerical responses of predators to cyclic lemming abundance: Effects on loss of goose nests. Canadian Journal of Zoology 79:525 – 532.